2010 nuclear forensics summer program report
TRANSCRIPT
2010 LLNL Nuclear Forensics Summer Program
Glenn T. Seaborg Institute Lawrence Livermore National Laboratory Physical
and Life Sciences Livermore, CA 94550, USA
Director: Annie Kersting ([email protected]) Education Coordinator: Nancy Hutcheon
Administrator: Camille Vandermeer Website https://seaborg.llnl.gov
Sponsors:
National Technical Nuclear Forensics Center, Domestic Nuclear Detection Office, Department of Homeland Security
LLNL: Glenn T. Seaborg Institute, Physical and Life Sciences Directorate
LLNL-AR-450120
2
The Lawrence Livermore National Laboratory (LLNL) Nuclear Forensics Summer Program is designed to give both undergraduate and graduate students an opportunity to come to LLNL for 8-10 weeks during the summer for a hands-on research experience. Students conduct research under the supervision of a staff scientist, attend a weekly lecture series, interact with other students, and present their work in poster format at the end of the program. Students also have the opportunity to participate in LLNL facility tours (e.g. National Ignition Facility, Center of Accelerator Mass-spectrometry) to gain a better understanding of the multi-disciplinary science that is on-going at LLNL.
Currently called the Nuclear Forensics Summer Program, this program began
ten years ago as the Actinide Sciences Summer Program. The program is run within the Glenn T. Seaborg Institute in the Physical and Life Sciences Directorate at LLNL. The goal of Nuclear Forensics Summer Program is to facilitate the training of the next generation of nuclear scientists and engineers to solve critical national security problems in the field of nuclear forensics. We select students who are majoring in physics, chemistry, nuclear engineering, chemical engineering and environmental sciences. Students engage in research projects in the disciplines of actinide and radiochemistry, isotopic analysis, radiation detection, and nuclear engineering in order to strengthen the ‘pipeline’ for future scientific disciplines critical to DHS (DNDO), NNSA.
This is a competitive program with over 200 applicants for the 8-10 slots
available. Students come highly recommended from universities all over the country. For example, this year we hosted students from UC Davis, Cal State San Louis Obispo, Univ. of Nevada, Univ. of Wyoming, Northwestern, Univ. of New Mexico and Arizona State Univ (See Table 1). We advertise with mailers and email to physics, engineering, geochemistry and chemistry departments throughout the U.S. We also host students for a day at LLNL who are participating in the D.O.E. sponsored “Summer School in Nuclear Chemistry” course held at San Jose State University and have recruited from this program.
This year students conducted research on such diverse topics as: isotopic
fingerprinting, statistical modeling in nuclear forensics, uranium analysis for nuclear forensics, environmental radiochemistry, radiation detector materials development, coincidence counting methods, nuclear chemistry, and actinide separations chemistry.
In addition to hands on training, students attend a weekly lecture series on
topics applicable to the field of nuclear forensics (see Table 2). Speakers are experts from both within and external to LLNL. Speakers are able to discuss the importance of their work in the context of advances in the field of nuclear forensics.
Graduate students are invited to return for a second year at their mentor’s
discretion. For the top graduate students in our program, we encourage the
3
continuation of research collaboration between graduate student, faculty advisor and laboratory scientists. This creates a successful pipeline of top quality students from universities across the U.S. Since 2002, 29 summer students have continued to conduct their graduate research at LLNL, 9 have become postdoctoral fellows, and 9 have been hired as career scientists. Two summer students have gone on to work at other national laboratories.
A big factor in the success of this program is the dedication of the staff
scientists who volunteer to mentor the summer students. In FY10, funding from The Nuclear Forensics Graduate Mentoring Program (sponsor: DNDO) helped to partially support the time staff took to teach the summer interns. Staff scientists could take the necessary time to develop an appropriate summer project, oversee the safety training and dedicate more time helping the interns maximize their productivity.
4
2010 Summer Students at Work
5
6
Table 1 Summer Student Roster 2010
Student Major University Year
Cameron Bates Nuc. Engineering UC Berkeley Grad
Megan Bennett Radiochemistry Univ. Nevada Las Vegas Grad
Greg Brennecka Geochemistry Arizona State Univ. Grad
Meghali Chopra Chemistry Stanford Undergrad
Eric Dieck Hydrology Illinois State University Grad
Steven Horne Nuc. Engineering U Texas Grad
Marianna Mao Math/Physics Harvard Undergrad
Christopher Matthews Nuc. Engineer Oregon State University Grad
Jon Oiler Geochemistry Arizona State University Grad
Matthew Snow Radiochemistry Washington State University
Grad
Table 2 Seminar Schedule 2010
Date Speaker Title of Presentation
June 24
Brian Powell, Clemson university Characterization of Actinide Speciation in Natural Systems
July 1 Ken Moody, LLNL Forensic Radiochemistry
July 8 Kim Knight, Julie Gostic, Ruth Kips, LLNL
Post- doctoral Research in Nuclear Forensics at LLNL
July 15 Sue Clark, Washington State University
Environmental Radiochemistry
July 22 Brad Esser, LLNL Isotope Hydrology
Aug 5 Jay Davis, former director of DTRA
Preparing for the Experiment one Hopes Never to do
Aug 12 Poster Session
8
Gamma-ray Intensity Ratio Determination of 235U
Megan E. Bennett, Nuclear Forensics Internship Program
Dawn A. Shaughnessy, Roger A. Henderson, Ken J. Moody, CSD, PLS
Post #This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Background
A requirement of nuclear forensics, environmental radiochemistry,
nuclear physics and the nuclear power industry is the need for
quantitative identification of 235U. One method used to achieve this is
gamma spectroscopy. To date, the energy and intensity ratio values for 235U !-ray energies above 300keV are inconsistent. The cause of this
inconsistence is most likely due to summing effects in the detector. It is
the goal of this study to resolve this issue and report the true intensity
ratios for 235U.
Method
A 235U source on Al was made using a self-deposition technique. A self-
deposited source of 235U was made on Al from an ~60mM 235U solution
at pH3. The source was then transported to the nuclear counting facility
at LLNL, where all measurements were made on a 30% High Purity
Germanium (HPGe) spectrometer. The sample was counted at various
distance and for various count times. The count times were chosen
such that an adequate number of counts under the peaks of interest
resulted. Each spectrum then underwent data analysis, as described
below.
Analysis
The raw data was processed with GAMANAL. The peak area in the
peaks of interest were then divided by the peak area of the 185.7keV
peak. This data was then plotted as a function of peak energy. The
error was based on counting statistics and calculated using standard
error propagation methods. The error in peak energy was smaller than
the data symbol. In order to determine if a peak would be observed, a
ratio of a strong, known peak and the peak of interest was taken.
Detector efficiency, the given distance and emission probability were
taken into account.
Results
Conclusions
In the !-ray spectrum of 235U 24 lines reported in the Table of the Isotopes (TOI) have been confirmed, while 28 other
peaks, above 200keV, reported by TOI remained unobserved. Of these 28 peaks, 11 peaks should have been
observed over the duration of the data collection based on the reported intensity values in TOI. The data presented
here indicates a serious discrepancy in not only the energy lines reported in TOI but also the intensity values.
Determination of intensity values of the peaks seen in this is underway.
Acknowledgements
The authors would like to acknowledge Todd Wooddy and Cindy
Conrado of the nuclear counting facility at LLNL
235U Peaks Matching TOI Reported Peaks
72.6 140.8 182.5 205.4 241.0
74.8 142.8 185.8 215.3 246.9
95.8 143.8 195.0 221.5 275.4
98.3 150.9 199.0 228.8 346.0
109.1 163.4 202.2 233.5
No %Ig reported !"#$%&$'()*(+(,(#-./0$( Should *1"#$%&$(+,22(314*5#
Possibly observed Should* not be observed Observed in all but 1 sample
Remaining TOI Reported Peaks
251.5 282.92 301.741 345.4 390.3 517.2
266.45 289.56 310.69 356.03 410.29 742.2
275.129 291.2 317.062 368.5 433.0 794.7
279.50 291.65 325.8 371.8 448.40
281.441 294.3 343.5 387.82 455.41
* This is based on reported %Ig in the Table of the Isotopes and calculated detector efficiency at a given energy and distance
9
The CANDU core is made up of 380 individual pressure tubes each with 12 half meternaturally enriched fuel bundles. For the simulation, each pressure tube is modeled as a singleassembly and assigned the average bundle power in that channel. A full core channel averagepower map was interpolated from a small data set produced by the reactor !"#$%&!$'( safetysimulations. The total core was blocked into eight regions of similar power and flux, andinterpolated over 100 days.
The power data was used as input into the TRITON module in the SCALE lattice cell code todetermine flux and fuel depletion using the 44-group ENDF/B-V library. Combined withmicroscopic cross section data, the fission rate and antineutrino rate was calculated for eachisotope.
After an initial drop that may an artifact from the simulation code, the total antineutrino ratedecreases linearly. As the U-235 content decreases and Pu-239 content increases, theantineutrino rate drops due to the 25% difference in detectable antineutrino emissionsbetween the two isotopes. The antineutrino rate closely follows the core isotopic content. ThePu-240 content reaches 7% of total plutonium content (weapons grade) on FPD 50.
! Roughly six antineutrinos per fission are emitted during beta decay of fission products andtheir daughters. The slight difference in antineutrino spectrum between fissionable isotopescauses a changing total antineutrino rate as the fuel is depleted. The time evolution of theantineutrino rate can determine if the reactor is on or off, calculate thermal power of the core,and predict isotopic content.
! The Point Lepreau Generating Station (PLGS) CANDU-6 reactor is undergoing a completereplacement of pressure tubes to extend the lifetime of the plant. The reactor will restart with anearly fresh core, an event that only occurs twice in )*+,-'( lifetime. The detector to bedeployed at PLGS will measure the antineutrino rate during the transition from a startup core(no refueling, fresh fuel, decreasing fission rate) to equilibrium core (online refueling, freshand depleted fuel, constant fission rate).
! The simulation presented here aims to predict the antineutrino rate in the first 100 days,prior to the start of online refueling.
!"#$"%&#'$"()'*#%)+(')#,%)+$'-#).//)+&00)1(2%')3*4-
5('%)6(3%0$"7)1'(8%--
!"#$"%&#'$"()%9$--$(")1%')+$--$(":'(8%--)(+)!"#$"%&#'$"();9$--$("
Antineutrino Flux Calculation for a CANDU Startup
The antineutrino flux emitted from reactors can provide a means of external safeguards monitoring. A computer model was created to simulate the initial 100 full power
days of a CANDU reactor prior to online refueling. These simulations will complement a detector to be deployed at the recently refurbished Point Lepreau CANDU reactor.
Using a time dependent power map the SCALE lattice code was utilized to calculate isotopic fission rates and depletion of the fuel. The total antineutrino rate can be shown
to decrease linearly after the initial startup as burnup of the fuel increases.
Topher Matthews, Nuclear Forensics Internship
Adam Bernstein, Physics Division, PLS
<=)>%-&0#-?
@=)6%#,(3-?
This model shows how the antineutrino rate is highly dependent on core
composition. While it does not take into account online refueling, it is still
applicable until full power day 70-80.
The lower content of Pu-240 present in the fuel before full power day 50
is more desirable for nuclear weapons. A potential proliferator mightreplace several or all fuel channels with fresh fuel and extract the
plutonium from the irradiated bundles before the percentage of Pu-240
increases to levels requiring more sophisticated separation techniques.
Since the antineutrino production is closely correlated to the fuel isotopic
composition, a jump in antineutrino counts could indicate early
replacement of fuel bundles. More simulations need to be performed to
predict the ability to detect this and several other diversion scenarios.
!"#$%&'()%*+(,'(-+.%/0.+(%1"+%2/$*#3+$%',%1"+%4565%7+*2(1-+01%',%80+(9:%;:%<2&(+03+%<#=+(-'(+%>21#'02?%<2;'(21'(:%/0.+(%@'01(231 78AB@CDAEF>BDFGHH5
!"!
!"#
$"!
$"#
! % & ' (
)*+,*-.+/,*01-*-/23145-67
89%:#
89%:(
;.9%:<
.=)A"#'(3&8#$("?
!
$!
%!
:!
&!
! #! $!!
5=+>?5
@A5
B.CC1;0D-/1EF3G
!B7 5,*""%0)1(2%')C4)7'(&1)
,**-/ H,IIC- 0.+-/
E-+-J+,0*1+>/-G>0CI
D(#*0)8('%)$-(#(1$8)8("#%"#
!"!
!"%
!"&
!"'
! #! $!!
5-+/,J1@0**-
B.CC1;0D-/1EF3G
89%:#
;.9%:<
;.9%&!
!"#$%&
E=)F$-8&--$("
!"!!
!"%#
!"#!
!"K#
$"!!
! #! $!!
L0/H
FC,M-I1+01+0+FC1/F+-1F+1B;E!
B.CC1;0D-/1EF3G
89%:#
89%:(
;.9%:<
+0+FC
10
Neptunium Sorption to Goethite
Mathew Snow, Nuclear Forensics Internship Program
Pihong Zhao, Mavrik Zavarin, and Annie Kersting, Chemical Sciences Division, PLS
Post # This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Results
Sorption curve for Neptunium to goethite •! Sampling performed on days 1, 4, and 8 after addition of neptunium (represented in
blue, green, and red respectively)
•! Data follows a linear trend for neptunium concentrations between 10-18 to 10-10 M
•! Begins to deviate from linearity at concentrations higher than 10-10 M
•! Approaches surface saturation at ~10-5 M.
Rate Dependence of Sorption •!Equilibrium reached within first day
•!Spiked blank samples showed ~15% activity loss
due to sorption on container walls
•!Minor increases in concentration in certain
samples as a result of changes in pH
Kd as a function of percent surface loading
•!Kd relatively constant from 10-10 to 10-4% site occupation
•!Decreases exponentially at concentrations greater than 10-4%
Preliminary Sorption Modeling
•!Concentrations <10-10 M aqueous Np (<10-4% site occupation)
support a Langmuir behavior
•!Concentrations >10-10 M aqueous Np (>10-4% site occupation)
support a Freundlich behavior
High Concentrations
(10-8 to 10-5 M)
Mid-Range Concentrations
(10-14 to 10-9 M)
Low Concentrations
(10-18 to 10-15 M)
Introduction
Large quantities of neptunium have been released into the
environment as a result of nuclear weapons programs and
reactor discharge.1 With a half life on the order of two million
years, coupled with its in-growth from parent nuclide
Am-241, Np-237 also represents a major contributors to the
long term (~100,000+ yrs) dose of proposed geological
repositories.2
Iron oxides, such as goethite (FeOOH), are commonly found
in rocks and soils. While the degree to which these minerals
(ad)sorb radionuclides has wide implications on their ability
to retard actinide migration, strong sorption also has the
potential to increase actinide mobility via colloid-facilitated
transport.3 Thus, an increased understanding of neptunium-
goethite interactions is essential for reliable transport
models, as well as possible methods for remediation and
retardation.
Background
While Np(IV) and Np(VI) have been observed in highly
reductive and oxic environments, respectively, Np(V) is the
predominant oxidation state under typical environmental
conditions.4 Most laboratory studies have analyzed
neptunium at much higher concentrations (>10-6 M) than are
frequently observed in the environment (10-12 M Np observed
at the Nevada Test Site), making the assumption that the
processes observed in the laboratory are simply scalable to
environmentally relevant concentrations.5,6,7
While neptunium-237 is the predominant neptunium isotope
found in the environment, its relatively low specific activity
makes it difficult to detect at lower concentrations (<10-7 M)
using nuclear counting techniques. Thus, for this study a
mixed approach was adopted, using Np-239 (half life of
2.345 days) for lower concentrations (10-18 to 10-15 M), a
239/237 mixture, in which Np-239 is used as a tracer, for
medium range concentrations (10-14 to 10-9 M), and pure
Np-237 for analysis of higher concentrations (10-8 to 10-5 M).
Methods
:
Introduction
Np-239 Purification
•!Milked from Am-243
•!1 mL,50-100 mesh AG BioRad column
•!50:1 HCl:HI load solution
•!Am-243 collected in the effluent, purified
Np-239 eluted using 1M HCl + 0.05M HF
Np-237 Purification
•!Purification from Pa-233 daughter
•!2 mL 100-200 mesh AG BioRad
column, concentrated HCl load solution
•!Pa-233 eluted using 9M HCl + 0.05 M HF
•!Np-237 eluted using 6M HCl +0.05 M HF
•!Repeated procedure until >99% pure
solutions obtained (determined by LSC
alpha/beta discrimination and gamma
counting)
Np(V) Characterization
•!Solutions reconstituted in 5M HNO3
to fix the Np oxidation state at (V)
and (VI)
•!Np was oxidized to (VI) in
concentrated HNO3 then reduced to
(V) in 5 M HNO3 using H2O2
•!Oxidation state was verified using
UV-Vis for Np-237 stock solutions.
UTEVA column was used to confirm
Np(V) for Np-239 spike solutions.
Sorption Study
•!18.8 m2/g goethite prepared using
established procedure7
•!Goethite suspended in a solution
of 5mM NaCl + 0.7mM NaHCO3
with 18! MQ H2O
•!pH maintained at pH 8.0+/- 0.5
•!Solutions constantly mixed using
rotary mixer
•!Sampling performed approximately
1, 4, and 8 days after Np addition
•!Samples were centrifuged at 3000
RPM for 2hrs to ensure adequate
solid-liquid separation
•!Np-239 and/or Np-237 in the
supernatants were analyzed using
LSC, with selected sample
verifications using gamma counting
and ICP-MS
References:
1-Smith, D.K., Finnegan, D.L., and Bowen, S.M., 2003. Journal of Environmental Radioactivity 67, 35-51.
2- Turner, D.R; Bertetti, F.P.; Pabalan, R.T. Soil Science Society of Amercia: Madison, WI, 2002; pp211-252 3- Kersting AB., D. W. Efurd, D. L. Finnegan, D. J. Rokop, D. K. Smith, J. L. and Thompson, (1999). Nature
397, 56-59
4- Dozol M, Hagemann R:.Pure Appl Chem 1993 , 65:1081-1102
5- Arai, Y., Moran, P.B., Honeyman, B.D., and Davis, J.A., 2007. Environmental Science & Technology 41,
3940-3944. 6-Nakayama, S. and Sakamoto, Y., 1991. Radiochimica Acta 52-3, 153-157.
7-Turner, D.R., Pabalan, R.T., and Bertetti, F.P., 1998. Clays and Clay Minerals 46, 256-269.
8-Schwertmann, U. and Cornell, R. M. (1991 VCH Publishers Inc., New York, NY.
Objectives
•!Study the Np-goethite sorption isotherm over a
wide concentration range (10-18 to 10-5 M)
•!Analyze Kd (equilibrium constant) as a function of
aqueous concentration
•!Implement novel approach using short lived Np-239
tracers to extend the sorption isotherm to
concentrations that are comparable to or lower
than Np concentrations observed in the
environment
1E-19
1E-18
1E-17
1E-16
1E-15
1E-14
1E-13
1E-12
1E-11
1E-10
1E-09
1E-08
1E-07
1E-06
1E-05
0 2 4 6 8
Aq
ueo
us C
on
cen
trati
on
(m
ol/L
)
Time (days)
Conclusion
•!Np-goethite sorption reaches equilibrium within 1 day
•!Kd relatively constant between 10-18 and 10-10 M (10-10
to 10-4% surface loading) then decreases exponentially
from 10-10 to 10-6 M
•!Np-239 tracers found to extend detection limits down
to ~10-18 M concentrations
•!Future work includes improving modeling capabilities using 2-site Langmuir and similar approaches to elicit
more information on Np-goethite sorption site affinity
Objectives
Conclusion
Methods
Results
References
1E+00
1E+01
1E+02
1E+03
1E
-18
1E
-17
1E
-16
1E
-15
1E
-14
1E
-13
1E
-12
1E
-11
1E
-10
1E
-09
1E
-08
1E
-07
1E
-06
1E
-05
1E
-04
Kd
(m
L/g
)
Aqueous Concentration (mol/L)
Experimental
Langmuir
Freundlich
1E+03
1E+04
1E+05
1E+06
1E
-10
%
1E
-09
%
1E
-08
%
1E
-07
%
1E
-06
%
1E
-05
%
1E
-04
%
1E
-03
%
1E
-02
%
1E
-01
%
1E
+0
0%
1E
+0
1%
1E
+0
2%
Kd
(m
L/g
)
Percent Surface Loading
Sampling 1
Sampling 2
Sampling 3
1E-17
1E-16
1E-15
1E-14
1E-13
1E-12
1E-11
1E-10
1E-09
1E-08
1E-07
1E-06
1E-05
1E-04
1E-19 1E-18 1E-17 1E-16 1E-15 1E-14 1E-13 1E-12 1E-11 1E-10 1E-09 1E-08 1E-07 1E-06 1E-05
So
rbed
Co
ncen
trati
on
(m
ol/g
)
Aqueous Concentration (mol/L)
Sampling 1
Sampling 2
Sampling 3
Day 1
Day 4
Day 8
Background
Day 1
Day 4
Day 8
11
Design of a Compton Vetoed Ge Detector
For Direct Measurement of Pu in Spent Nuclear Fuel
Cameron Bates, Nuclear Forensics Internship Program
Stephan Friedrich, Physics - Physical and Life Sciences
Post # This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
A typical 1-GWe power plant discharges up to 30,000 kg
of spent fuel per year. Of that roughly 300 kg is Pu. Even though it is a small amount of the total mass of spent fuel
it is enough for 30 nuclear devices(~10 kg per device). This becomes an issue when the fuel is reprocessed,
because the Pu is separated from highly radioactive
fission products. If it is not possible to determine how much Pu was in the fuel prior to dissolution, there is no
way to verify that none was diverted.
Direct Pu measurement is
currently impossible
Spent fuel spectrum1: Pu is masked by Compton
scattering from Cs and other fission products
Plutonium?
Approach: Thin HPGe detector
with a Compton veto behind it
References: 1.Passive Nondestructive Assay of Nuclear Materials. Office of Nuclear Regulatory Research, 1991.
Results: Prototype setup
•!30x30x8mm3
planar Ge detector
•!8” Pb collimator
•!2xCsI veto:
!!3”Øx6”
!!6”Ø(outer)x3”
•!Pulse tube cooler Pb
Cryostat
CsI Veto Ge
•! Veto rejects low angle Compton scatter into region of interest
•! HPGe has the energy resolution to resolve low energy Pu X-rays •! Ge and scintillator geometry optimized to improve veto
•! Cs-137 used for system characterization •! Scintillator material chosen to optimize rejection
Optimization of scintillator
geometry: as large as possible
Scintillator material choice:
maximize stopping power
Simulated spent fuel spectra
Scatter angle "= 90
CsI veto
BGO veto
No veto
NaI veto
Photopeak Cs-137
Veto thickness
= 4cm
12cm
Veto
radius
= 5cm
7cm
Photopeak Cs-137
No veto
Compton background Compton background
Motivation:
Where is the plutonium?
239Pu
241Pu
Pu X-rays
U X-rays
238Pu
154Eu
144Ce
103Ru
Will it work?
A spent fuel source was
modeled in MCNP5 using results from a benchmarked
ORIGEN simulation. The results show that it is
theoretically possible to directly
measure Pu with this system. Tests of this prototype and
further modeling will prove this.
Pb collimator Scintillator veto Ge
Cryostat
"
thickness
rad
ius Source
12
35S: New Tracer of Groundwater Travel Time Near MAR OperationsStephanie Diaz, Nuclear Forensics Internship Program
Bradley Esser, Richard Bibby, Chemical Sciences Division, PLS
Post #This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
This work was completed with the help of J. Moran, S. Roberts, M. Holtz, E. Dieck from LLNL, and B. Chong from the Water
Replenishment District of Southern California.
Acknowledgements
Moran, J., and M.S. Halliwell, 2003. Characterizing Groundwater Recharge: A Comprehensive Isotopic Approach. Awwa Research Foundation.
McDermott, J., D. Avisar, T.A., Johnson, and J.F. Clark, 2008. J. Hydrologic Engrg.,13: 1021-1028.
Michel, R.L., D. Campbell, D. Clow, and J.T. Turk, 2000. Water Resour. Res., 36: 27-36.
References
Model of MAR Operations
Assumptions:1) Constant 35S flux
2) 35S behaves conservatively
3) Homogenous 35S and SO4
distribution in pond
4) No mixing with non-MAR
water
Surface RunoffWell
35SPond = Ao > 0 mBq/L
35SWell = At
! Managed artificial recharge (MAR) is a water management tool designed to
augment water supplies.
! Current CA regulation requires the location of municipal wells to be 150 m
from MAR unless a six month travel time can be determined.
! Dating MAR water has previously involved extensive field work (gas
tracers: SF6, Xe) or utilizing isotopes with longer lived half lives.
! 35S is a cosmogenically produced radionuclide (t !=87 days) that has been
useful for studying recharge in alpine basins, but its application in MAR
studies has been limited due to high SO4 concentrations.
Project Goals:1) Refine method for natural waters containing high SO4
2) Evaluate potential of 35S method to estimate groundwater travel times at
two MAR locations: Rio Hondo Spreading Grounds (RHSG) and Alameda
County Water District (ACWD) (Fig 1)
Introduction
ACWD
RHSG
Figure 1. Location of MAR study sites and RHSG recharge pond.
Monitoring Well
Production Well
Rio Hondo Spreading Grounds
Well SF6 (days)* 35S (days)
MW1
Alameda County Water District
Well 124Xe
(days)**
136Xe
(days)**
35S (days)
Bunting
Kaiser
Samples collected and processed based on modified techniques from Michel
et al., (2000):
1) Process approximately 1000 mg SO4 per sample by acidifying and suspending
Amberlite ion exchange resin in sample for 2 hrs
2) Elute SO4 from resin with NaCl solution
3) Add BaCl2 to precipitate BaSO4
4) Suspend BaSO4 in Instagel for liquid scintillation counting
Method Development for Natural Waters
To refine the method for higher SO4 MAR waters, the following variables
were examined: pH effect on SO4 adsorption, sample exposure time to resin
(Fig A), total amount of resin (Fig B), intermittent mixing of resin in sample
(Fig C), and effects of heat, HNO3, H2O2 on color quenching (Fig D).
Analytical Methods
C
1 2 3 4
Method Refinement! Varying the pH (pH 1 to pH 7) and total amount of resin (15 g to 25 g) had
little effect on SO4 adsorption
! Continued mixing of resin in sample was more effective in adsorbing SO4
relative to intermittent mixing.
! Exposing the BaSO4 precipitate to heat and HNO3, or HNO3 and H2O2
resulted in a higher tSIE values, suggesting that these treatments reduce
the quench associated with a sample.
Travel Time EstimatesTable 1. Comparison of travel time determined by 35S to previous tracer studies.
*McDermott et al. (2008) **Moran and Halliwell (2003)
Results
WellDate
CollectedSO4
(mg/L)
35S (mBq/L)
35S Travel Time (days)
Tracer Travel Time (days)
Tracer
RHSG-Pond 2 June 2010 184.0 8.76 NA NA NA
RHSG-MW1 22 Apr 2010 105.7 3.60 75 46 SF6*
RHGS-MW1 23 May 2010 146.0 2.89 89 46 SF6*
ACWD-B Pond 14 Jul 2010 86.5 2.76 NA NA NA
ACWD-B 14 Jul 2010 79.9 0.04 MDA 56 124Xe**
ACWD-K Pond 14 Jul 2010 90.7 3.86 NA NA NA
ACWD-K 14 Jul 2010 93.3 0.74 173-235 <20 136Xe **
Intermittent Mixing
A BResin Exposure Time Varying Amount of Resin
With the exception of one monitoring well in ACWD, 35S activity in the ponds
and wells was above MDA. As expected based on the simplified model, 35S
activity was higher in MAR surface water from the ponds relative to the
groundwater from the wells. The older travel times resulting from the 35S
method relative to previous studies may be due to mixing or water-rock
interactions that are not accounted for by this simplified model. These
preliminary results demonstrate the potential use of 35S as a tracer in MAR
operations.
Future Studies
! Further comparison of age dated MAR wells with varying travel times using 35S method
! Determine seasonality of 35S activity in MAR surface and groundwater
! Determine potential stratification of 35S and SO4 within MAR ponds
Conclusions
74 270 2791317886
tSIE Values
1 L Del Valle Samples20 L Del Valle Samples
Heat + HNO3
Heat + HNO3 +
H2O2
D
Heat
t = Travel Time
( )!""""""""""A tln(2) Ao
t = ln
Cosmic
RaysCosmic
Co
sm
ic
13
Uranium and Plutonium Isotope Detection Using Coincidence Methods
Steven Horne, Nuclear Forensics Internship Program, University of Texas
Dr. Tzu-Fang Wang, Chemical Sciences Division, PLS, Lawrence Livermore National Lab
Post # 424640
Release Number: LLNL-POST-446391
This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
ABSTRACT In this project, we aim to look at uranium and plutonium samples using coincidence counting methods. The main isotopes of interest were uranium-235, uranium-238, plutonium-239, and plutonium-240. As a result of this project, we
were able to determine the isotopic ratios for our premade samples using coincidence techniques, which could play a key role in nuclear non-proliferation and accountability.
INTRODUCTION
•! Can we take a uranium or plutonium sample and figure out its isotopic ratios with a high amount of background
present? •! High resolution gamma ray spectrometry has been in use for over 50 years and is useful for element and isotope
detection in environmental, biological, and radioactive samples.
•! Using multiple detectors, one has the capability of looking at isotopes with cascading decays without interference from the Compton continuum of fission products with high energy gamma rays.
•! Using these techniques, we can determine uranium or plutonium isotope ratios, even in the presence of fission products.
RESULTS
METHODS We used a total of three high purity germanium detectors. The two coaxial detectors are useful for looking at
gamma rays of high energy (100 – 1800 keV), and the planar detector is useful for looking at gamma rays of low energy (20 – 300 keV). Using a list mode data acquisition system we are able to detect gamma-rays from the
samples with a coaxial – coaxial coincidence and a coaxial – planar coincidence configurations. By analyzing energies and time of the data among the detectors, we are able to determine the isotopic ratios using the ratios of
the energy gated coincidence events. We verified this technique using three known isotopic compositions of
uranium gamma-ray standards consisting of .5%, 5%, and 50% U-235.
DISCUSSION When attempting to determine the concentration of U-235 in a sample, one often looks at the 186 keV peak in the spectrum. This peak is often interfered with by the Compton continuum produced by long-lived fission products.
In order to overcome this barrier, we utilize the coincidence counting method, with which we compare the ratios of
the 186 – 202 keV coincidence peaks to the 1001 keV single peak, which is of high enough energy to not be affected by the Compton continuum. From this method, we can develop an equation to predict the concentration of
U-235 in an unknown sample. With the results achieved, we can estimate the equation of best fit to be y =x/(.254+x), where x is the ratio of 186 x 202 keV coincidences to 1001 keV single counts and y is the concentration of
U-235 in the sample. This estimation does take into consideration the dead time of the counts. In order to fully
demonstrate the effectiveness of the method, we need to look at the angular dependence of the cascading emissions, as well as the efficiencies of the detectors in use.
U-235 Concentration
186 x 202 Coincidence :: 1001 keV Ratio
.5% 8.46e-4 ± 3.71e-4
5% 8.08e-3 ± 6.47e-3
50% .255 ± .014
0.0001
0.001
0.01
0.1
1
10
0 20 40 60 80 100 120
Counti
ng R
ati
o
U-235 Concentration (%)
Concentration of U-235 by 186 x 202 keV Coincidence
Best Fit Line
Observed Data
The above figure shows the three detectors in use – on the left and right are the two coaxial detectors, and on the bottom is the
planar detector.
The above graph is an example of a coaxial - coaxial coincidence spectrum being analyzed. Red and green spots correspond to a lower
number, while the blue and white spots correspond to more coincidences. The axes correspond to the channels of the respective detectors. This
spectrum came from the 50% U-235 sample.
14
Separation and isotopic analysis of activation products for nuclear forensics Meghali Chopra, Nuclear Forensics Internship Program!!
Aaron Glimme, Department of Energy Academies Creating Teacher Scientists!
Scott Tumey, Atmospheric, Earth and Energy Division, PLS!
Abstract: We separated and prepared various metal samples for isotopic characterization by accelerator mass spectrometry (AMS) including aluminum, iron, nickel and
cobalt. Separations were performed using column chromatography and selective precipitation. Initial concentration measurements were performed using inductively
coupled plasma optical emission spectroscopy (ICP-OES) and the performance of various AMS cathode matrix metals was evaluated with high purity iron samples.!
This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.!
Introduction: !Nuclear forensics plays a crucial role in the deterrence of nuclear terrorism.
Transition metal activation products can provide valuable information about the
components and method of delivery of a nuclear device. Unfortunately the
prevalence of large background quantities of metals such as aluminum, iron, nickel,
and cobalt mean that the activation products of transition metals must be measured
against a large background signal. Isotopes of iron, cobalt and nickel have stable
isobars that can make measurements particularly difficult. Two possible approaches
to these problems are improved detector discrimination and the use of chemical
differences to separate the isobars before analysis. Our focus has been on chemical
differences that will allow the determination of transition metal isotope ratios with
unparalleled accuracy in this challenging sample environment.!
Methodology: !
Discussion: !
Implications: !
Typically concentrated nitric acid would be used to dissolve
metal samples but strong passivation effects with transition
metals necessitates the use of more dilute 2-8M nitric acid. !
In preparation for AMS samples are further cleaned by selective
precipitation, then dried and oxidized in a furnace. Finally
samples are packed in metallic cathodes.!
The concentrations of known standard solutions were first measured using ICP-
OES. After a linear calibration curve had been established, the three unknowns
were measured. Based on the initial calibration, sample 1 had a concentration of
2.613±0.008 ppm, sample 2 had a concentration of 2.439±0.005 ppm, and the
process blank had a concentration of 0.0477 ±0.005 ppm. !
Results: !
Iron samples were analyzed by AMS to determine the effects of different
cathode materials. High purity iron samples were loaded into cast aluminum
and pressed powder matrixes of niobium, tantalum, tungsten, and titanium.
!These samples were run on the AMS tuned to count mass 59 and measure
! !the current of mass 56 isotopes. !
Additional work remains for the development of a complete separation and analysis
protocol. First, more analysis of the purity that can be achieved by our combination
of column and precipitation separation is needed. Future experiments will include
ICP measurements of iron, cobalt, and nickel solutions as well as AMS
measurements of cobalt and nickel in their aluminum, niobium, tantalum, tungsten
and titanium matrices. Second, the new AMS detector designed for the separation
of iron, cobalt, and nickel isobars must be validated. After the sample preparation
process has been refined and the AMS detector has been optimized, it will be
possible to very quickly characterize the activation products for iron, cobalt and
nickel from real world backgrounds with unprecedented sensitivity. !
A key component to achieving accurate AMS isotopic measurements is the
minimization of unwanted isobars. For sample preparation, this means chemically
separating the elements of interest from those that interfere with the AMS
measurements. This is particularly difficult with iron, cobalt and nickel because they
have very similar chemical properties. Typical separations developed for geological
applications are time consuming. Rapid analysis is a critical requirement for
forensics, so our focus has been aimed at developing less time-intensive separation
methods.!
Our initial results indicate the most commonly
used cathode materials are all well suited to use
with iron AMS analysis. Outliers in the data can
be attributed to a skewed normalization as the
live ratio begins to drop. Tests for optimal nickel
and cobalt matrices are also ongoing.!ICP-OES
results for aluminum were consistent with initial
estimates based on the size of the aluminum
samples originally digested.!
One item of note is the slightly elevated!
concentration of the digestion process blank. More analysis of the materials and
process used in the initial digestion is needed to ensure that no additional aluminum
is being inadvertently added during the digestion. Efforts to accurately detect iron,
!cobalt, and nickel via ICP are ongoing and will constitute a vital factor in ensuring
! !the completeness of the chemical separation. !
""#"$%&'($))*+,-!!
15
Uranium ore concentrate (UOC), or “yellowcake”, is the !nal product in the mining and milling of uranium ore, and represents an intermediate step in the utilization of uranium for its energy potential.Yellowcake is a fungible commodity that is commonly traded worldwide, but is also a regulated nuclear material. In cases of illicit tra"cking, chemical impurities and isotopic signatures imparted on the UOC by the
source rock and/or mining process have the potential to allow the UOC to be traced to its point of origin. High precision isotopic ratios from a few key elements can provide invaluable clues to the provenance of illict material.
Geolocation of Nuclear Materials
UCRL-POST-233253
!"#!$
!"#!%
!"#&'
!"#&(
!"#'!
!"#'$
!"#
$%!&#
$
!"#!)
!"#&!
!"#&)
!"#'!
!"#')
!")&&$ !")&&( !")&&% !")&'! !")&'' !")&'$
!"#
$%!&#
$
'()*+%'((*+
Greg Brennecka - Nuclear Forensics Internship Program Lars Borg - Chemical Science Division, PLS
This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344
*+!"#!#,
-./+++-.%# %(
!"#!#,+01
!"#!%$
!"#!%$+01
!"#&!!2+!"#&!!
AustraliaCanada
NamibiaCzech Rep.
USA
Determining of the provenance of a sample of UOC requires a variety of techiques and methods. Certain isotopic systems, such as Sr and Nd, have proven be particularly useful in determining thetype of rock from which nuclear material was derived, a critical step in the geolocation of a sampleof unknown origin.
The isotopic data obtained from UOC can be used to producesingle-element maps of a region, as
shown in this bedrock Sr isotope map of Great Britian (redrawn from NERC
Isotope Geosciences Laboratory).
This information can be combinedwith isotopic data from other elements
to greatly reduce the possible sourceregions of UOC. This permits
idenifacation of a unique isotopic !ngerprint for many UOCproducing regions of the
world. An example ofthis is shown usingSr and Nd isotopic
values, illustrating !veunique signatures.
The Strontium (Sr) and Neodymium (Nd)isotopic composition of uranium ore is dictatedby 1) the composition inherited at the time of ore formationand 2) radiogenic ingrowth from Rb and Sm parent isotopes. Thus, Sr and Nd isotopiccompositions of UOC represent a long-term averageparent/daughter ratio (i.e., Rb/ Sr; Sm/ Nd)of the rocks in the region from which UOC is derived.Chemical separation of Sr and Nd from the rest of the elements in UOC to enable precise measurementof isotopic ratios for each element for use ingeolocation.
Sr Nd
Sr isotopes values from UOC are shown, representing a few of thehundreds of UOC producing regions from around the world. Knowledge of only one isotope ratio does not produce unique valuesfor each region, but can greatly narrow the search.
Australia
CanadaNamibia
Czech Rep.
USA
87
87
147
147
87 143
16
Nuclear forensic analysis can help determine the provenance of interdicted or declared nuclear material by identifying material characteristics (signatures)
particular to its geographic or processing origins. Regression models (classifiers) calibrated on an existing database of known materials extract the
comparative multivariate signatures that most differ from one source (class) to another. Based on these signatures, unknown samples are then assigned to a
class in the database with some degree of confidence. Normally, signatures are inferred from classifiers trained on multiple samples per class. However, in
the case of nuclear material, often only one sample per class can be obtained. I assess the performance of such single-sample models.
For nuclear safeguards, non-proliferation, or law-enforcement
applications, nuclear forensic analysis may be used to
determine or verify the declared origins of uranium ore
concentrate (UOC). Multivariate signature analysis (e.g.,
trace element concentrations) can be used to help determine
the origins of a given UOC sample through comparison of the
unknown sample to a database of known samples. However,
since the database contains only one sample per class for
some classes, models built on this database are weakened by the lack of information on in-
class scatter. In this project, I characterize the performance of models calibrated on single
samples.
UOC (yellowcake) sample.
LLNL-POST-446054 This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Methods
A classifier’s performance depends on both its sensitivity
(the number of true positives correctly identified) and
specificity (the number of true negatives correctly
identified). As the decision boundary is made stricter, the
sensitivity of the classifier decreases while the specificity
increases. This relationship is illustrated by the ROC curve,
which plots sensitivity and specificity as the boundary is
changed. The area under the ROC curve can be used to
compare different models and to gauge confidence interval
outputs.
ROC curves cannot be calculated for a model based on single-sample data because in-class
variance can only be inferred from a distribution of samples. Hence, we split a dataset with
multiple samples per class into a single-sample calibration set and a validation set and
manually calculated estimated ROC curves based on the validation data.
Although the single-sample model’s area of 0.6 was low compared to the near-1
areas of models with multiple samples per class, the limited amount of data made
for a conservative estimate of the area, while the calculation for the model with
multiple samples is a relatively optimistic one. Overall, the calculation shows that
models based on single samples are still viable tools for nuclear forensic analysis.
The ROC area was calculated by varying the decision boundary based on the y-
predicted intervals from the validation data. After obtaining a discrete number of points
on the ROC curve, the area was computed by the trapezoid method. The average area
for a single-sample model was 0.6.
Marianna Mao, Nuclear Forensics Internship Program
Martin Robel, Chemical Sciences Division, PLS
Estimated ROC Curve from Single Sample Model
A Class 1 estimated ROC curve calculated from the dataset on
the left. Points were generated by incrementally moving the
Class 1 decision threshold and tallying results at each step.
Leave All But One Out Validation (LABOO)
Class 1 Modeled from Sample 1 Only
Yellowcake dataset split into calibration (dark) and validation (light)
data. Calibration data were used to construct the model. Validation
data are samples from known sources treated as unknowns. All
Class 1 samples above the threshold are true positives. Samples
from other classes above the threshold are false positives.
Sample 28: 55 Sample 29: 20
17
Development and Characterization of 10B Containing Thermal
Neutron Capturing Materials
J. Oiler1, Nuclear Forensics Internship Program
L.F. Voss2, A.M. Conway2, Q. Shao2, C.E. Reinhardt2, R.T. Graff2, D.M. Estrada2, R.J. Nikolic2
Post #
This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
1. Arizona State University; Tempe, AZ, 85287
2. Lawrence Livermore National Laboratory; Livermore, CA, 94550
ABSTRACTDevelopment of a solid-state thermal neutron detector is on-going at LLNL. The detector incorporates silicon pillar PIN diodes at high aspect ratios with the trenches in between the pillars filled with 10B.
The summer project was to expand on the existing work in two important areas:
1) Design, build, and carryout a thermal neutron particle experiment to directly compare different boron capturing compounds
2) Develop a fabrication process for a detector incorporating the neutron capturing material boron oxide .
The thermal neutron experiment has been designed, the components have been delivered and are currently being set up and tested. A basic fabrication process for the boron oxide detector is completed,
but optimization is still on-going.
BORON OXIDEElemental boron is an excellent material for solid-state thermal neutron detectors due to very high
thermal neutron cross-section and its ability to be used in MEMS fabrication processes. However,
we wanted to see whether isotopically enhanced boron oxide could also be used as a thermal
neutron capturing material because boron oxide is a cheap material that requires non-vacuum
based processing. We have had success with the initial boron oxide fabrication and process
optimization is under way before device testing.
The boron oxide device fabrication process described below is the same as the boron device
fabrication in steps 1, 2, 5.
1) Spin and pattern photoresist with pillar geometry on a wafer with a PIN Si diode structure.
2) Use DRIE to etch 3-D pillar geometry into substrate.
3) Pipette a saturated solution of enriched 10-boron oxide powder (Ceradyne Boron Products,LLC) and methanol over the entire wafer. Bake the wafer at 500 !C to allow the boron oxide to
melt and flow into the etched trenches and anneal (A).
4) Polished back overfilled boron oxide glass using a slurry of oil and alumina particles (B).
5) Deposit aluminum electrodes on the top and bottom of the of the structure.
Boron oxide overfill
Boron oxide Boron oxideSi pillars Si pillars
A B
Detection of thermal neutrons is important in the fields of
astronomy, physics, and homeland security. Presently, thermal
neutron detection is carried out using helium-3 tubes, but
these devices are costly, power hungry, large in size, and not
robust. Therefore, moving to the use of solid-state thermal
neutron detectors is desirable. Small, portable, low power, and
robust sensors would allow for easier screening of shipping
containers and airport luggage for hazardous radioactive
materials.
The solid-state thermal neutron detector researched at LLNL
uses a 3-D high aspect ratio pillar structure. The silicon is
doped to be a PIN diode and undergoes deep reactive ion
etching (DRIE) to form pillar structures (A). Next, the gaps
between the pillars are filled with 10B using chemical vapor
deposition (B) (work done at University of Nebraska-Lincoln).
After removing the excess B, Al electrodes are deposited on
top and bottom of the device (C) for electronic readout.
A B C
INTRODUCTION
There are a number of solid-state detectors incorporating
different boron compounds2-4. Each of these detectors may
use a different boron compound (e.g. boron, boron carbide,
boron nitride, boron phosphide). Furthermore, each of these
materials may be deposited using multiple methods. These
materials and depositional methods may result in different
neutron sensitivities.
BORON COMPOUND SENSITIVITY EXPERIMENT
The capture of a thermal neutron causes a 10B atom (thermal neutron cross-section of 3387 barns) results in an immediate
decay via the following reaction1:
10B+n ! 7Li + !
In 94% of reactions the Li and ! particles have energies of 0.84 MeV and 1.47 MeV, respectively. In the other 6%, the Li
and ! particles have energies of 1.01 MeV and 1.78 MeV. The "#$#%$&'() sensitivity to thermal neutrons will be directly
related to the density of the 10B within the thermal neutron capturing boron compound.
Directly comparing different detectors is a challenge for the following reasons:
1) Design and fabrication of each of these detectors may be different; they can have different size sensing areas,
geometries, noise levels, amount of material, etc.
2) Sensitivity of a device or material in one lab may not match results in another lab with the same test setup. Because
thermal neutrons will scatter off many materials and be absorbed by a host of others, neutrons scattered from all
directions (not simply just in a direct line from the source) will be delivered to the detector. Therefore, the geometry,
structure, and setup of the entire lab will play a role into the directional flux of thermal neutrons .
3) Fabricating working devices for material comparisons can be very time consuming.
For these reasons, we are constructing an experiment where we can compare different boron containing materials and
boron deposition processes to one another by depositing the material on silicon, thereby eliminating issues of using
different devices and saving time on device fabrication .
Experiment setup diagram
Alpha particle detector
Silicon Boron compound
REFERENCES1) Knoll, G.F., Radiation Detection and Measurement, 3rd Ed. (Wiley, New
York, 2000).
2) Nikolic, R.J., Conway, A.M., Reinhardt, C.E., Graff, R.T., Wang, T.F., Deo,
N., and Cheung, C.L., Applied Physics Letters, 93, 133502, 2008.
3) Osberg, K., Schemm, N., Balkir, S., Brand, J.I., Hallbeck, S., Dowben,
P.A., Hoffman, M.W., IEEE Sensors Journal, 6, 1531, 2006.
4) Uher, J., Pospisil, S., Linhart, V., Schieber, M., Applied Physics Letters,
90, 124101, 2007.
25 µm
7
Disclaimer This document was prepared as an account of work sponsored by an agency of the United States government. Neither the United States government nor Lawrence Livermore National Security, LLC, nor any of their employees makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States government or Lawrence Livermore National Security, LLC. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States government or Lawrence Livermore National Security, LLC, and shall not be used for advertising or product endorsement purposes. Auspices Lawrence Livermore National Laboratory is operated by Lawrence Livermore National Security, LLC, for the U.S. Department of Energy, National Nuclear Security Administration under Contract DE-AC52-07NA27344.